xlvL PREFATORY REMARKS 



that the balance of evidence is in favour of the ganoid affiid* 

 ties [of the Cephalaspidse] ; but I do think that there art i 

 sufficient grounds for dividing the ganoids of Agassiz, and 

 adopting M 'Coy's term Placostei,* or Owen's Placo-Ganoidei. 

 T prefer the latter, as Placostei is used by the Germans as 

 synonymous with Placoid&i. In this case, the Cephalaspides 

 would form a family of the placo-ganoid sub-order." 



Whether we abide by Agassiz's arrangement, 'nodified by 

 Professor Owen's subdivisions, or, with Professor Huxley, 

 desire to adopt a more recondite and exclusive style of classi- 

 fication, for whose gradual development we must patiently 

 wait until farther researches and clearer evidence will have 

 brought it to light, — in any case it is worthy of remark, 

 that Professor Huxley in no way endeavours to exclude the 

 Cephalaspidse from the true ganoids, in order to degrade them, 

 but precisely the contrary. On this point his opinion is un- 

 doubtedly entitled to great weight Let us hear what he 



" "Without doubt, there is a singularly close resemblance 

 in the structure of the dermal plates, between Cephalaspis and 

 Megalichthys, — the last being very probably a true ganoid ; 

 but the point of diflference is noteworthy ; — it is precisely the 

 characteristic ganoid layer which is absent in Cephalaspis. 

 On the other hand, the arrangement of the hard tissues in 

 Pteraspis reminds one almost as strongly of Ostracion, an 

 undoubted teleostean. The existing fishes to which Cepha- 

 laspis presents the nearest resemblance in form, — viz., Lori- 

 caria and Callichthys, — are Siluroid teleosteans, and not 



• See the latter part of Note C, p. 309. 



